Could this be combined with Theo Jansen's Strandbeest idea, to make a machine that (very) slowly melts a glass road through the desert, powered only by the sun and the wind?
I wonder what possible applications could come out of using a fresnel lens+a stirling engine to drive pistons in a desert:
https://youtu.be/CDCTANU8Tfk?t=1m1s
The goal of such a system would be self-replication[1]. So the material that the stirling engine is made from should be sourced locally from the environment. Perhaps iron or aluminium.
Also make use of molten salt energy storage systems to keep the system running over night.
Calculate your doubling time and extrapolate to determine when you've solved the worlds energy problems, or when you've turned it into grey goo. ;)
The key to this kind of a system is to reduce both part and element count.
That way full-closure of hte system will be easier to achieve and full self-replication will be an easy step to take.
Stirling engines and Fresnel lenses are the right path for both electrical and thermal energy production. If you can keep as much of the system as possible derived from sand and clay it's pretty much producable anywhere on Earth.
This will be so awesome, once the thing can integrate lenses, solardriven circuitry and silica-gel to its constructions.
Imagine a structure that collects dew and heat during the day, and transports that water up during the night. Providing a canopy and a microclimate in the dessert from which forests can sprout. And its printed on side, with no additional materials needed.
Eh, usually use up of limited resources for irrigation (oil, fossile water), but this could create transforming machines from glass and energy. Well.. if you like sand boarding or dune buggy this thing has a downside..
[1] http://www.exploratorium.edu/strandbeest/meet-the-beests